Induced radiative techniques such as Raman and fluorescence detection have become useful tools in evaluating the properties of materials. Not only are such processes useful in determining unknown constituents, but the techniques are becoming increasingly popular in process-control situations to monitor the presence or absence of chemical precursors, intermediate products, and so forth.
To characterize a composition in a remote or hostile environment, optical fibers may advantageously be used to deliver excitation energy to a sample under investigation and to carry emitted radiation back to means for spectral analysis. An excitation source path may take the form of a laser providing a stimulus at a nominal wavelength coupled to an input fiber, and a collection path maybe made up of a second fiber carrying return radiative information to a spectral analysis tool such as a spectrograph.
Such remote spectral analysis presents technical challenges, however, including the strong scattering signature of the material used for the optical fiber, this interference potentially being generated by both the laser excitation in the illumination fiber and any strong Rayleigh (unshifted) scattering allowed to enter the collection fiber. These spurious fiber signatures can compete with, or even overshadow, the desired signature of the sample under test, particularly when long lengths of fiber are used.
In a typical system of this kind, a source of excitation is directed onto or into a substance under evaluation, be it a solid, liquid or gas, and the wavelengths emitted therefrom are analyzed in terms of their shift relative to the excitation wavelength. As some forms of detection, such as Raman detection, rely upon the discrimination of relatively weak signals, a notch filter is ordinarily included prior to detection to remove the excitation wavelength indicative of Rayleigh scattering so as not to overwhelm the much weaker shifted wavelengths of interest.
Although such systems operate upon a wavelength shift as opposed to an absolute value at one or more frequencies, it is still nevertheless important that such systems be calibrated in order to obtain an accurate reading. Automated internal wavelength calibration is ordinarily carried out in spectroscopic systems by physically moving a source of known spectral content into the field of view of the system. This normally blocks the external measurement bypass temporarily in order to execute the calibration procedure.
External wavelength calibration through the same light collection optics as used in sample measurement always provides the most accurate possible wavelength calibration. Internal calibration methods normally compromise accuracy to some extent due to slight differences in the intensity profile of a spectral line image at the detector or detector array.
Laser wavelength tracking and subsequent Raman shift correction are not normally carried out in Raman systems. The operating wavelength of the excitation laser is normally assumed to be accurately known, and Raman shifts of a measured spectrum computed accordingly. In addition, the high-efficiency laser blocking filter used to eliminate the much stronger Rayleigh scatter from the spectrometer, tends to render external injection of a laser tracking signal or calibrating reference emission lines near the laser wavelength problematic. This is unfortunate, since the operating wavelength of many solid-state lasers can change from day to day by an amount that can degrade the accuracy of Raman shift calculations by an amount significant to many spectroscopists.